The COVID-19 pandemic has changed many things in business, producing a new normal that all of us now operate in—and analytics is no exception.
“As companies adapt to the new normal created by COVID, one of the primary questions we’re asked in analytics is how to retrain artificial intelligence (AI) models with a more diverse data set,” said David Tareen, director of AI and analytics at SAS.
Tareen said that the pandemic altered the integrity of the analytic data models that many companies had in place. This required companies to redefine and retrain those models so the analytics could address new business and environmental conditions that the pandemic brought about.
SEE: How COVID-19 is disrupting the enterprise and what you can do about it (TechRepublic Premium)
In some cases, existing data models that had been so reliable suddenly began to underperform to the point that they needed tweaking. In other cases, companies lacked data and applications altogether to deal with the COVID crisis, and they had to find ways to obtain the data they needed and develop new analytics models quickly.
“in one case, airports had been using predictive modeling to understand and improve aircraft traffic flow,” Tareen said. “What they found with the pandemic was that these models had to be retrained and additional data sources added before the models could start accurately predicting the new normal traffic pattern that resulted from COVID.”
In some instances, new types of analytics data and applications had to rapidly be created to deal with the COVID crisis. The east Indian state of Odisha found itself in this situation.
Odisha wanted tools to measure the aggressiveness of the pandemic and determine how the pandemic was spreading. It also wanted analytics that could help it assess what the patient loads would be on healthcare facilities like hospitals and clinics, and how it could best manage quarantines and other measures needed to combat the advance of COVID. Finally, it wanted risk assessment tools.
The problem was that Odisha didn’t have the data or the tools that were needed to do the job.
The on-the-ground issues Odisha was facing were critical. There were bottlenecks in mobilizing people and deployment of infrastructure due to the strict rules that were in place for the country lockdown. Logistics, road transport, trains, police personnel availability and resources at numerous government departments were some of the biggest hurdles.
To help meet the challenge, Odisha brought in SAS to help in a “ground floor” analytics project. “While the overall project delivery took four to six months, Odisha’s dashboard was built in only six weeks,” said Kunal Aman, head of marketing for SAS India. “Odisha used analytics for monitoring the virus transmission, infection rate, and state healthcare infrastructure readiness.”
SEE: Snowflake data warehouse platform: A cheat sheet (free PDF) (TechRepublic download)
What were the lessons learned from Odisha and the airports?
1. It’s necessary to roll with the changes
In any crisis (in this case, COVID) the underlying assumptions and fundamentals of analytics can radically change, rendering them ineffective. Companies must develop agile techniques to continuously refine and recalibrate their analytics so the analytics stay in step with what is happening on the ground.
2. You may need to ask for help
You can find yourself in a situation where you don’t have the internal analytics tooling or expertise to rapidly respond to a crisis at all. In these cases, assistance from knowledgeable vendor or consultant can help.
3. Analytics are essential for organizations to function
Finally, the COVID crisis may well be the inflection point that transformed analytics into a mission-critical capability. Without up-to-date analytics that can inform and actualize incident response, there are few ways for organizations to know what is going on in their environments at any point in time, how to assess risk or what to do.
This is how COVID transformed analytics from a “nice to have” to a “must have” corporate capability.